Abstract: The laminas and natural fractures play a crucial role in controlling the hydraulic fracturing of shale reservoirs. Accurate characterization of the reservoir rock mass structure is essential for optimizing the stimulated reservoir volume. This study focuses on the laminas and natural fractures of the Pingdiquan Formation hydrocarbon source rocks in the northeastern margin of the Junggar Basin, constructing a detailed three-dimensional geological structure model. Initially, we obtained high-resolution image data of the vertical profile of the Pingdiquan Formation hydrocarbon source rocks through close aerial photography by an unmanned aerial vehicle(UAV) and digital imaging. Subsequently, we constructed a high-precision three-dimensional digital model of the profile based on the image data to identify and interpret the natural fractures and laminar structures. Combining the parameters of natural fractures and laminas measured manually on-site, we determined three main sets of high-angle fractures through cluster analysis, with fracture length following a lognormal distribution. The average thickness of the laminas was about 0.57 m, with the overall lamina thickness following a negative exponential probability distribution. Based on the statistical characteristics of natural fractures and laminas, by utilizing Monte-Carlo simulation techniques, we constructed models for natural fractures and laminas, respectively. Finally, we overlaid these models to generate a three-dimensional rock mass structure model incorporating both natural fractures and laminas and verified by the measured profile. This study introduces UAV technology to acquire high-resolution image data of steep-profiled hydrocarbon source rocks and interprets structural information on natural fractures and laminas, enhancing the refinement of rock mass structure characterization. The findings provide a more reliable three-dimensional model for numerical or physical simulations of the Pingdiquan Formation hydrocarbon source rocks in the Junggar Basin.